Embed Size (px)
Citation preview
1
USING A LIFE COURSE APPROACH TO UNDERSTAND INVOLVEMENT IN VOLUNTEERING IN AUSTRALIA
Authors
Edith Gray, Siew-Ean Khoo, and Anna Reimondos Australian Demographic and Social Research Institute, The Australian National University
This paper is based on Stage 1 of a research project funded by the Department of Families, Housing, Community Services and Indigenous Affairs. Stage 1 focuses on the analysis of HILDA data to examine volunteering over the life course. Stage 2 examines volunteering for different types of organizations using Negotiating the Life Course data. We thank the staff of the Research and Analysis branch of FaHCSIA for their comments on the paper. The data used for this research come from the Households Income and Labour Dynamics in Australia (HILDA) survey funded by the Commonwealth Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA) and conducted by the Melbourne Institute for Applied Economic and Social Research (MIAESR). The research findings are the product of the researchers and the views expressed do not necessarily represent the views of MIAESR or the Minister for Families, Housing, Community Services and Indigenous Affairs and cannot be taken in any way as expressions of Government policy.
2
INTRODUCTION
Volunteers make an important contribution to Australian society. There has been a long history of volunteering in Australia, and in recent years there has been an increase in the percentage of people involved in volunteering (ABS, 2008; FaHCSIA, 2008). In 2006 the percentage of the Australian adult population that volunteered was 35% (ABS, 2008). These high levels of participation contribute to economic and social benefits for Australian society, and there is considerable evidence that the contributions volunteers make are beneficial for the volunteer as well as those they help (Li and Ferraro, 2006; Musick, Herzog, and House, 1999; Thoits and Hewitt, 2001).
The definition of ‘volunteer’ is in some ways problematic as there is debate over whether it should be distinguished from time given to caring activities and similar ‘informal’ activities. However, it is widely acknowledged that volunteering is ‘time that is given freely’, and is thought of as being ‘public’ (Wilson 2000, pp216). For many purposes the use of ‘formal volunteering’ (e.g. Volunteering Australia 2005) is used to express the act of providing time to not-for-profit, charitable and community organizations.
In this paper we examine formal volunteering in Australia. We focus specifically on how a life-course perspective can benefit the study of people’s involvement in volunteering. Most of the quantitative research on volunteers in Australia has looked cross-sectionally at the characteristics of volunteers, so we know who is most likely to be involved in volunteering, but we know little about movement in and out of volunteering and the factors which are associated with these movements. We focus our paper on changes in people’s lives and how these changes affect volunteering.
BACKGROUND
While much is known about the demographic characteristics of volunteers in Australia, there is less knowledge about how changes in people’s lives affect volunteering or about the factors that are associated with the starting or stopping of volunteering.
Current knowledge about the characteristics of volunteers in Australia (ABS, 2008) shows the following groups more likely to volunteer: (1) women; (2) employed men; (3) part-time employed women; (4) healthy individuals; (5) people who live outside capital cities; (6) people with young children and (7) people aged over 55.
While studies that provide information on the types of people who are involved in volunteering are essential for understanding patterns of volunteerism, they do not explicitly investigate how life course changes are related to volunteering. There is a growing amount of research which proposes that simply looking at individual characteristics and resources such as socio-economic resources, time, health, gender and ethnicity or race as indicators of voluntary involvement is not
3
sufficient (Rotolo, 2000; Tang, 2006; van Willigen 2000). The life course perspective is an extremely useful framework for conceptualizing the factors that are associated with the dynamics of volunteering, as it specifically takes changing roles and life events as central to understanding behaviour.
Life course stage and volunteering
The life course perspective has five general principles: life-span development, or studying lives over time which provides rich information; agency as the construction of lives and interaction with social and structural constraints; time and place emphasizing the historical context matters; timing which suggests that when events occur in the life course matters, and, linked lives which highlight interaction with ‘important others’ (Elder et al., 2004).
An important concept in understanding life course research is that age is an ‘empty indicator’ (Musick and Wilson, 2008:221). By this, Musick and Wilson mean that age is often used as a proxy for life stage, but that it is not a person’s age per se that is associated with behaviour, it is the circumstances which are associated with a person at that age which are important. In using a life course perspective, we need to understand the dominant life course stages, and investigate how movements between these stages are associated with changes in the behaviour of interest: which in this case is volunteering.
Young adulthood
The first life course stage (after childhood and adolescence), is that of young adulthood. Young adulthood is generally represented by those in their late teens to late twenties or early thirties. The rate of volunteering in this age group is relatively stable across this age range, with around 30 per cent of young adults volunteering at ages 18-24 and 25-34 (FaHCSIA, 2008). However, there is some variation by age group in terms of the median hours volunteered. Those aged 18-24 have higher median hours volunteered (about 50 hours per year) compared with the 25-34 year age group (40 hours per year). Further, for young men there is little difference in volunteering hours across the two age groups, but for women there is a considerable drop at ages 25-34.
If we compare these levels of participation to other ages groups, young adults are less likely to volunteer and they volunteer for fewer hours than people in middle adulthood; they are also more likely to volunteer but volunteer for fewer hours than people in later adulthood (FaHCSIA, 2008).
The life course perspective emphasizes time in individual lives. By focussing life course stage we are able to consider the other important events and conditions that individuals experience. Hence, there are times of people’s lives where there is little time available or there are many demands. At these times people may be less likely to volunteer which is consistent with role overload theory (Rotolo, 2000; Wilson, 2000). In considering how this applies to young people who are transitioning to adulthood; this time of life has been described as ‘demographically dense’ (Rindfuss, 1991): a time when multiple transitions are being made such as leaving education,
4
starting work, leaving home, starting relationships and having families, etc. (Oesterle, et al., 2004; Stoker and Jennings, 1995). These types of events should be considered when investigating involvement in volunteering.
Middle adulthood
Middle adulthood is characterized as a more stable period of life (Musick and Wilson 2008). During these years many people have paid work and a large proportion have children many of whom are of school age. As Musick and Wilson found for the U.S., this is a time in which ‘the volunteer rate peaks’ (2008: 222). This is true also for Australia. There are however significant differences between men and women. For women aged 35-44 almost half participated in volunteering in 2006, dropping to just under 40% of women aged 45-54. For men, the rate was almost 40% in both age groups (FaHCSIA, 2008). While the median hours for women was relatively stable in both age groups (around 55 hours), for men the median hours was less in the 35-44 age group (about 35 hours compared to 55 hours for men aged 45-54).
These participation patterns reflect the life course stage of men and women. Men at these ages are much more likely to be involved in full-time work, while women are likely to have part-time labour force participation at these ages. Women are also more likely to have primary responsibility for childrearing activities, including volunteering-related school activities.
For people with partners another factor associated with the likelihood of volunteering may be whether their partner volunteers. Life course theory recognizes that it is important to consider the effect that ‘significant others’ may have on behaviour. In the case of volunteering, there is some evidence that whether an individual volunteers is related to whether their husband or wife volunteers (Rotolo and Wilson, 2006). They also found that wives had more influence on their husband’s volunteering than vice versa.
Later adulthood
There are many life changes that occur in later ages that may play a role in understanding participation in volunteering. An important life course transition is retirement from employment. However, later adulthood is not only associated with retirement; changes in health and widowhood are also important transitions that may affect volunteerism (Butrica, et al., 2009; Hank and Erlinghagen, 2009; Tang, 2006). Phyllis Moen refers to the ‘mid course’ years, which draw a distinction between people in the early stages of later life (their 50s, 60s and early 70s) and those at the oldest ages. Typically, these years are characterized by changes in role repertoire: often from a greater intensity in paid work (Moen and Fields, 2002). Goss (1999) also noted the very different participation rates in volunteering as people reached the ‘oldest old’ ages, which is related to health problems.
These two stages of later adulthood are evident in Australia. Participation rates in volunteering decline steadily from age 55. However, the hours volunteered are actually highest in the years 65-
5
84. So while those in later adulthood are less likely to volunteer, those who do volunteer contribute many hours. The drop in both participation and hours at the oldest ages is considerable.
People’s lives are dynamic: their ability and availability for volunteering changes depending on their situation. The life course approach is a useful framework for understanding the affect of transitions on volunteering.
DATA AND METHOD
Data
Analysis is based on the Household, Income and Labour Dynamics in Australia (HILDA) data. HILDA is a prospective panel study which is funded by the Australian Government through the Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA). HILDA collects a large amount of participant information, including economic and subjective well-being, income, labour, and family dynamics (Wooden and Watson, 2001).
Panel participants are aged 15+. They are surveyed annually, and this paper uses the data which are currently available (2001–2008). In 2008 over 7,000 households involving 12,785 individual interviews were conducted (MIAESR, 2010). We use a question on volunteering which was asked in the self-completion questionnaire. Participation in the self-completion questionnaire is slightly lower than overall participation (N=11,187 valid responses for volunteering).
There is a single question in HILDA about involvement in volunteering, and the question is asked as part of a larger set of questions relating to time spent on different types of activities in a ‘typical week’.1 We note that other researchers have expressed concern about the way questions on volunteering are asked (Hall, 2001).2 However, as we focus on individual response and change over time, we are less concerned with measuring propensity to volunteer which is obviously better captured by other collections (e.g. ABS Voluntary Work Survey).
The question asks:
How much time would you spend on each of the following activities in a typical week?
…Volunteer or charity work (for example canteen work at the local school, unpaid work for a community club or organisation
Those who spent at least some time volunteering are coded as having volunteering, while non-volunteers are those who have written ‘0’ in the hours spent volunteering. 1 Please see Appendix 1 for a discussion on the HILDA question. 2 According to Hall (2001) the quality of any questionnaire on volunteering depends on how well it (1) ensures that respondents understand what is being asked; (2) contains questions that respondents are able to answer, and (3) maximizes the ability of respondents to recall past giving and volunteering behaviours.
6
Method
We conduct two sets of analysis on volunteering. The following describes the cross-sectional and longitudinal techniques examined.
Cross-sectional analysis
Firstly, the 2008 data is used to look cross-sectionally at the data to describe the characteristics of those who volunteer. Crosstabulation and logistic regression techniques are used. The crosstabulation simply shows us the percentage that volunteers by certain socio-demographic characteristics. The logistic regression provides the opportunity to better understand the propensity to volunteer while controlling for other life factors.
The factors which are used to examine volunteering include: sex of respondent, age, country of birth, marital status, presence of children, caring responsibilities, education, employment status, income, self-rated health, region, religion and life satisfaction.
Longitudinal analysis Secondly, in order to examine how changes in the life course affect volunteering we use an event history technique known as a discrete-time multivariate hazard model using logistic regression. It provides the ability to analyse:
1. The starting of volunteering (characteristics of those who started volunteering, given that they were not volunteering the first time they were observed)
2. The cessation of volunteering (characteristics of those who stopped volunteering, given that they were volunteering the first time they were observed).
Analytical method for longitudinal analysis
The data are analysed using discrete-time event history analysis (using the logistic regression xtlogit command in STATA). This analytical technique was also used by Butrica, et al (2009) in their analysis of entries into and exits from formal volunteer activities among older Americans between 1996 and 2004.
The data analysis is conducted separately for the age groups 15-34, 35-54 and 55+, given that there are some life course events which are important for one group while irrelevant for the other. For example, having a birth is less likely for people aged 55 and over.
The strength of these models is that it allows the measurement of starting and stopping of volunteering, so we can better understand the factors that influence people’s decisions to start or stop volunteering. Using the life course perspective, we specifically look at how changes in people’s lives influence the starting and stopping of volunteering.
7
The analysis has been conducted separately by age to understand the life course factors that are associated with the different age groups. The three age groups which are modeled are: (1) 15–34, (2) 35–54, and (3) 55+.
A detailed description of the technical aspects of the data analysis is available from the authors.
Dependent and independent variables
The dependent variable is whether or not the respondent started (or stopped) volunteering. The independent variables include fixed variables – sex and country of birth; and time-varying variables (note that time varying variables represent variables where a respondent’s status may have changed since the previous wave) – age, region, education, equivalised income, marital status, employment, time, moved residence. Additionally, having a birth, and presence of children, are included in the model for respondents aged 15–54.
For both starting volunteering and stopping volunteering, three different models are run:
1. Total sample (all ages) 2. Younger sample (aged 15–34 the first time they were observed) 3. Middle-aged sample 35-54 the first time they were observed) 4. Older sample (aged 55+ the first time they were observed)
The following describes the variables used in the models, and variables that are specific to the younger sample are noted.
Fixed variables:
Sex
Country of birth 1. Australian 2. English-speaking country 3. Non-English speaking country
Time varying variables:
Age
Change in marital status. Derived from looking at marital status in every wave, and classifying any changes across two waves into five categories.
1. Married in wave T & also married in wave T-1 (i.e. married continuously since last wave)
2. Cohabiting in wave T & also cohabiting/married in wave T-1 (i.e. cohabiting continuously since last wave or cohabiting in previous wave and now married)
8
3. Single in wave T-1, but cohabiting/married in wave T (i.e. started a cohabitation or a marriage since last wave)
4. Cohabiting or married in wave T-1, but single in wave T (i.e. separated/divorced/ widowed from a cohabitation or marriage)
5. Single in wave T & wave T-1 (i.e. continuously single since last wave)
The effect of categories 3 and 4 (either starting or ending a cohabitation or marriage), compared to those who were continuously married was similar for both starting and breaking from volunteering. In the final models, categories 3 and 4 were therefore combined due to the small number of cases observed undergoing these two relationship transitions.
Highest education level 1. Bachelor degree or higher 2. Certificate/Diploma 3. Year 12 4. Less than Year 12
Employment transitions 1. Continuously working (PT or FT) 2. Continuously not working 3. Start working 4. Stop working
Both categories 3 and 4 were more likely to start and to stop volunteering than those who were continuously working (although for starting volunteering those who stopped working had a higher likelihood of volunteering than those who started working). Due to small number of cases starting and stopping work in the 55+ age group for the break from volunteering model, these two transition categories were combined in the final model.
Equivalised household income (in thirds)3 1. Bottom third 2. Middle 3. Top third
Health change Collapsed version of variable in self completion questionnaire which asks respondents to rate their health compared to one year ago
1. Improvement compared to a year ago
3 This was calculated according to the code outlined on page 5 of this HILDA training manual: http://www.melbourneinstitute.com/hilda/HILDA%20Training_practical%20lab_DW.pdf
9
2. About the same as a year ago 3. Worse now than a year ago
Caring commitments Based on the number of hours spent caring for someone from the self-completion questionnaire. A three way distinction was made between those who were not caring, those who were caring for less than ten hours a week, or for ten hours a week or more.
1. Not caring 2. Caring for <10 hours in a typical week 3. Caring for 10+ hours in a typical week
Moved residence in the past 12 months Dichotomous variable indicating whether the respondent has moved residence in the past 12 months
Geographic Location (Accessibility / Remoteness Index of Australia (ARIA)) 1. Major city 2. Inner Regional Australia 3. Outer regional or remote Australia
Time Controls for which wave this is for the respondent (the more time they are observed for the more likely they are to experience a stop in volunteering)
In the model for the younger sample, two additional variables were included.
Had a birth in the past 12 months 1. Had a birth/adoption in the past 12 months 2. Did not have a birth/adoption in the past 12 months
Child composition 1. No children at all 2. No own resident children aged under 14 3. Own resident children only under 5 4. Own resident children aged under 5 & 5–14 5. Own resident children aged 5–14 but not <5
In the final model those aged 15–54 with no children at all, and those with no own resident children aged under 14 were combined.
10
RESULTS
Characteristics of people who volunteer and who do not volunteer
The findings from the HILDA survey on the characteristics of people who volunteer (Table 1) are similar to those from other studies (ABS 2006; FaHCSIA 2005). The crosstabulation results show that the percentage volunteering is higher for (1) women than men, (2) married people than single people or people in de facto relationships, (3) older people than younger people, (4) people with children aged 5–14, (5) people with more education, (6) people working part-time or not in the labour force than people who are employed full time or unemployed, (7) people born in Australia or other English speaking country than migrants from non-English speaking countries4, (8) people who live outside the major cities, (9) people with better health and (10) people for whom religion is important.
Table 1 about here.
As some of these observed patterns between the demographic variables and volunteering rates could have been influenced by other variables, a logistic regression was conducted to control for the influence of these variables.
In this analysis the dependent variable was defined as: 1= volunteers, 0=does not volunteer. Independent variables are the same as those in Table 15. The model was run for the sample as a whole, as well as for men and women separately to test for any possible interaction effects with gender.
The results of logistic regression to control for the relationships between some of the demographic and social variables confirm the findings discussed above. They also show that there is a U-shaped relationship between caring and volunteering: those who have high amounts of caring or no caring for an older or disabled person are less likely to volunteer than people with moderate amounts of caring responsibilities. There is no significant difference between males and females and no association between volunteering and life satisfaction once other demographic and social characteristics are controlled for. The logistic regression results also show that the characteristics of people who volunteer and who do not volunteer are similar for both men and women (Table 2).
4 English language comprehension has been identified by people born in non-English speaking countries as the most common barrier to their participation in volunteering (Volunteering Australia 2007). It has also been suggested that one reason for the lower rate of volunteering by people born in non-English speaking countries may be that they do not identify the unpaid work they do in their communities as volunteering or that it fits the formal definition of volunteering (see discussion in Volunteering Australia 2007). 5. Religion was not included in the regression analyses because the question was asked in waves 4 and 7 but not in wave 8.
11
While the logistic regression results show no significant difference between men and women in their propensity to volunteer, the age pattern of volunteering varies between men and women. For women there is a peak in the late thirties to forties, which may be a time when they are involved with volunteering with various sports or recreation organisations for their school-aged children. Men are much more likely to volunteer from age 55.
Table 2 about here.
Life course transitions and starting and stopping volunteering
When respondents who are initially observed to be not volunteering are followed to see who start volunteering, it is not surprising to see that the characteristics of these people are similar to those who are identified earlier as more likely to volunteer. Both the descriptive (Tables available on request) and the event-history (longitudinal) analyses (Table 3) show that women, people aged 35–44, people with higher levels of education, people with children aged 5–14 and people living outside the major urban areas are more likely to start volunteering during any observation period. People born in non-English-speaking countries are less likely to start volunteering compared to people born in Australia or other English speaking countries.
When survey respondents who are volunteering are observed to stop volunteering, the characteristics of those who are more likely to stop are also similar to those people identified earlier as less likely to volunteer (Table 3). Men are more likely to stop than women, and younger people aged 15–34 as well as older people aged 55 and over are more likely to stop than people aged 35–44, even after controlling for any change in their marital or employment status. People with children under age 5 are also more likely to stop volunteering.
The data also show that many life course events are associated with starting or stopping volunteering. Further, there are life course events that differently affect the likelihood of starting or stopping volunteering depending on life course stage.
Table 3 about here.
Young adulthood
For example, for those in young adulthood starting a new live in relationship has no effect on the propensity to start volunteering. In fact at this life stage it is only people who are (ongoing) cohabiters that have a lower likelihood of starting to volunteer. However, in terms of stopping volunteering it is observed that young adults who marry/divorce, or start/end a cohabitation are more likely to stop volunteering. Like the other life course groups, young adults who are continuously working are the least likely to volunteer, however the odds of volunteering for those who have less (or less stable) attachment to the workplace does not show the large differences that are apparent in the later age groups. In contrast to the U-shaped results for caring which is
12
evident in the descriptive results, we see that particularly for the young cohort, that those who have caring responsibilities are much more likely to start volunteering, and are much less likely to stop volunteering, than those with no caring responsibilities. Perhaps this is due to being exposed to caring.
There appears to be no difference in the likelihood of volunteering for those from other non-English speaking countries as compared to the Australian born. This pattern is only evident in the young adulthood group and can be seen as a positive sign of involvement. There is much discussion about less involvement in organized volunteering by people from non-English speaking backgrounds—a pattern we do see in the other life course groups—but these results indicate that it is not across all age groups.
Moving and having a child in the last 12 months are both associated with a lower likelihood of starting volunteering, but are not significant for stopping volunteering. The coefficients are large (but not significant) for stopping volunteering, and we attribute this to the small number of observations. Starting volunteering is strongly associated with having school-aged children for this group.
Middle adulthood
In many ways this group looks similar to the young adults in terms of the patterns of starting and stopping volunteering due to life event. However there are some notable differences, often in the propensity to start or stop volunteering which distinguish their behavior from the young adults.
For example, the effects of changes in marital status are substantially different for this life stage. In terms of starting volunteering those who are continually married are much more likely to start volunteering. They are also less likely to stop volunteering. Those who are continuously employed are also substantially less likely to start volunteering than others and this effect is larger than for the young adults. It appears that employment is a very important part of the life course for this group.
Health concerns also have a negative effect on starting to volunteer, while moving has a significant effect on stopping volunteering. Again it is evident that the birth of a new child has a large negative effective on starting to volunteer but also a large positive effect on stopping. Having a school-aged child is associated with starting volunteering.
Later adulthood
Among people aged 55 and over, a change in marital status such as the start or the end of marriage or a de facto relationship is associated with a greater likelihood of starting volunteering. There is a strong effect of moving on stopping volunteering for this life course group. Long caring hours is associated with a lesser propensity to start volunteering and a greater likelihood of stopping.
13
CONCLUSION AND FURTHER RESEARCH
This paper contributes to our understanding of starting and stopping volunteering in the Australian context. Most research simply looks at the propensity to volunteer by considering individual resources and characteristics. We extend the scope of quantitative research on volunteering by adding a life course dimension. This dimension allows us to consider the important effects of changes in people’s lives on starting and stopping volunteering.
We proposed that volunteering was associated with life course factors, and that changes in these factors were related to participating, and to starting and stopping volunteering. We also argued that there may be different factors that affect volunteering at different life stages, so some factors may be particularly important in the middle adult years, and other factors which may be more important in later adulthood and in retirement. Consistent with this proposition we found that some factors were associated with a lower likelihood of starting to volunteer (and a higher likely to stop volunteering). These include: continued employment, declining health, moving home, and having a birth. Further, there were different patterns by life course stage of the respondent. For example, a change in relationship had a large positive relationship with starting to volunteer for older people, while for those aged less than 55 it had an effect on stopping volunteering. If a static approach is taken, the observation of these types of patterns is not possible.
A more detailed understanding of the effect of life course stage on volunteering can be investigated by examining the different types of volunteering activities that are associated with different stages of the life course; the amount of time spent on these activities; and how changes in the life course can affect the amount of time spent on volunteering. While these cannot be addressed using HILDA, there are other datasets which may be suitable including the Negotiating the Life Course. This is being undertaken in the second stage of this project. A dynamic approach, like that used here, provides a much more complete understanding of involvement in volunteering.
14
REFERENCES
Australian Bureau of Statistics (ABS). (2008) Australian Social Trends 2008, Article: Voluntary Work. Cat. No. 4102.0. Canberra.
Butrica, B.A., Johnson, R.W., and Zedlewski, S.R. (2009) Volunteer dynamics of older Americans, Journal of Gerontology: Social Sciences, 64B(5), 644–655.
Elder, G.H., Johnson, M.K., & Crosnoe, R. (2004) The emergence and development of life course theory, In J.T. Mortimer and M.J. Shanahan (Eds) Handbook of the Life Course, pp.3–19.
Families, Housing, Community Services and Indigenous Affairs, Department of (FaHCSIA) (2008) Volunteering in Australia changing patterns in voluntary work 1995–2006, Canberra.
Goss, K. (1999) Volunteering and the long civic generation, Nonprofit and Voluntary Sector Quarterly, 28(4), 378–415.
Hall, M. (2001) Measurement issues in surveys of giving and volunteering and strategies applied in the design of Canada’s National Survey of Giving, Volunteering and Participating. Nonprofit and Voluntary Sector Quarterly, 30(3), 515–526.
Hank, K., & Erlinghagen, M. (2009) Dynamics of volunteering in older Europeans, The Gerontologist, 50(2), 170–178.
Li, Y. and Ferraro, K. (2006) Volunteering in middle and later life: Is health a benefit, barrier or both? Social Forces, 85(1), 497–519.
Melbourne Institute of Applied Economic and Social Research (MIAESR). (2010) HILDA Survey Annual Report 2009, University of Melbourne: Melbourne.
Moen, P. & Fields, V. (2002) Midcourse in the United States: Does unpaid community participation replace paid work? Ageing International, 27(3), 21–48.
Musick, M.A., Herzog, A.R., & House, J.S. (1999) Volunteering and mortality among older adults: Findings from a national sample, Journal of Gerontology: Social Sciences, 54B(3), S173–S180.
Musick, M.A. & Wilson, J. (2008) Volunteers: A Social Profile. Indiana University Press: Bloomington.
Oesterle, S., Johnson, M.K., & Mortimer, J.T. (2004) Volunteerism during the transition to adulthood: A life course perspective, Social Forces, 82(3), 1123–1149.
15
Rindfuss, R.R. (1991) The young adult years: Diversity, structural change, and fertility, Demography, 28(4), 493–512.
Rotolo, T. (2000) A time to join, a time to quit: The influence of life cycle transitions on voluntary association membership, Social Forces, 78(3), 1133–1161.
Rotolo, T. & Wilson, J. (2006) Substitute or complement? Spousal influence on volunteering, Journal of Marriage and Family, 68, 305–319.
Stoker, L. & Jennings, M.K. (1995) Life-cycle transitions and political participation: The case of marriage, American Political Science Review, 89, 421–433.
Tang, F. (2006) What resources are needed for volunteerism? A life course perspective, Journal of Applied Gerontology, 25(5), 375–390.
Thoits , P. A., & Hewitt, L. N. (2001) Volunteer work and well-being, Journal of Health and Social Behavior, 42(2), 115–131.
van Willigen, M. (2000) Differential benefits of volunteering across the life course, Journal of Gerontology: Social Sciences, 55B(5), S308–S318.
Volunteering Australia (2005) Definitions and principles of volunteering, Information Sheet, Available at: http://www.volunteeringaustralia.org/files/AOAL2F8K3S/VA%20Definitions%20and%20Principles%20June%202005.pdf
Volunteering Australia (2007) National Survey of Australian Volunteers of Diverse Cultural and Linguistic Backgrounds, Available at: http://www.volunteeringaustralia.org/files/BRTVP4DHB8/CALD%20National%20Survey%20final%20report%20with%20cover.pdf
Wooden, M., & Watson, N. (2001) The Household, Income and Labour Dynamics in Australia (HILDA) Survey: An Introduction to the Proposed Survey Design and Plan, HILDA Discussion Paper No. 1/00, University of Melbourne: Melbourne.
Wilson, J. (2000) Volunteering, Annual Review of Sociology, 26, 215–240.
16
Appendix 1
Questions on volunteering in HILDA
The type and number of questions on volunteering used in surveys can have a large impact on estimates of volunteering behaviour.
For example there is a wide range of survey estimates on the percentage volunteering in Australia, although it should be noted that these are not directly comparable given the different samples surveyed:
HILDA (2008): 18%
Negotiating the Life Course (2006): 53%
Giving in Australia survey6 (2005): 41%
Voluntary Work Survey (ABS 2006): 34%
According to Hall (2001:517-518), the quality of any questionnaire on volunteering depends on how well it:
ensures that respondents understand what is being asked
contains questions that respondents are able to answer
maximizes the ability of respondents to recall past giving and volunteering behaviours.
The following outlines how the question on volunteering is asked in HILDA, and the advantages or disadvantages of each questionnaire for studying volunteering.
HILDA questions
In HILDA, there is no separate question that asked only about volunteering. Instead the question on volunteering is asked as part of a larger set of questions relating to time spent on different types of activities in a ‘typical week’ (see Diagram 1).
Volunteers are identified as people who spend at least some time volunteering, while non-volunteers are those who have written in 0 in the hours spent volunteering.
Since the focus of this report is on individual response and change over time, the estimate of the percentage volunteering as measured in HILDA is not an issue
The question asks:
How much time would you spend on each of the following activities in a typical week?
…Volunteer or charity work (for example canteen work at the local school, unpaid work for a community club or organisation
Those who spent at least some time volunteering are coded as volunteers, while non-volunteers are those who have written ‘0’ in the hours spent volunteering.
6 This is a comprehensive survey on volunteering. Sample=6,000 respondents aged 18+ by telephone interview. The data is available for free from ASSDA following the necessary approval.
17
Diagram 1: HILDA question on volunteering
While this question is useful in getting a sense of time spent on different tasks, it is not ideal for getting an estimate of volunteering for several reasons:
Respondents are likely to be thinking of a typical week, and may therefore be prone to forget volunteering episodes which are less frequent e.g. take place once a month or less.
Related to this, there is no specific time period specified (e.g ‘in the last 12 months’). If the respondent typically volunteers for something which may be seasonal, such as particular sport only played in the summer months, they may leave out this volunteering because it is winter when they are answering the questionnaire, and therefore not in the ’typical week’ they have in mind at the time.
Although the question does contain a few specific examples of volunteering behaviour, some respondents may still be unclear if what they do fits in the category of ‘volunteer or charity work’.
Respondents who do not volunteer may not be properly following the instructions, and instead of writing ‘0’ as instructed, they may just leave the question blank leading to missing values which are then excluded from the analysis leading to an overestimation of volunteering.
In HILDA the question has been asked the same way for every wave from wave 1 to wave 8. The only difference is that in Wave 1 only hours were asked for, whereas in Waves 2 onwards it was hours and minutes. The percentage of respondents who were volunteering was relatively stable across the waves (Appendix Table 1).
18
Appendix Table 1: Percentage of respondents volunteering (valid responses - excluding missing, unweighted)
Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 %
volunteering 23% 19% 21% 20% 18% 19% 18% 18%
# of respondents 12,464 11,366 10,668 10,188 11,448 11,682 11,374 11,189
Source: HILDA 2001-2008.
Comparison of patterns of survey responses between HILDA and other data collections
The following figure (Appendix Figure 1) demonstrates the differences found between HILDA (2008), NLC (2006) as compared with the Voluntary Work Survey (which was collected as part of the Australian Bureau of Statistics General Social Survey 2006). The figure shows the percentage of respondents volunteering by age and sex. The pattern observed using HILDA data is different from that observed using both NLC and the Voluntary Work Survey, particularly at the younger ages. The volunteering rates from the ABS survey and NLC are considerably higher than those derived from HILDA, especially at the younger ages. A major reason for this difference is likely to be the difference in question wording.
Despite these differences in observed propensity to volunteer, the figure does show that the likelihood of volunteering varies by age, and most likely, life stage.
Appendix Figure 1: Percentage of respondents volunteering by age and sex, comparison of HILDA 2008, ABS General Social Survey 2006 & NLC 2006
Source: HILDA 2008 (Wave 8), ABS (2006), NLC (2006)
0
10
20
30
40
50
60
70
18-24 25-34 35-44 45-54 55-64 65-74 75-84
Per
cen
tag
e vo
lun
teer
ing
Age group
Males - NLC 2006
Females -NLC 2006
Males - ABS 2006
Females - ABS 2006
Females - HILDA 2008
Males - HILDA 2008
19
Table 1. Sample description and percentage of respondents who volunteered, by selected demographic characteristics, Wave 81
Sample
N Sample
% % who
volunteer
Sample N
Sample %
% who volunteer
Sex*
Equivalised income*
Male 5,190 46 15 Bottom third 3,630 32 19
Female 5,997 54 19 Middle 3,759 34 17
Top third 3,798 34 16
Age group*
15-24 2,061 18 9 Child composition*
25-34 1,635 15 12 No child at all 7,167 65 16
35-44 2,000 18 18 No children aged under 15 1,220 11 17
45-54 2,052 18 19 Children aged 0-4 806 7 8
55-64 1,565 14 19 Children aged 0-4 & 5-14 402 4 23
65+ 1,874 17 25 Children aged 5-14 1,496 13 25
Country of birth* Caring for someone*
Australia 8,990 81 18 No 10,291 92 16 English speaking
t1,089 10 19 Yes, <10 hours a week 615 6 32
Other 1,106 10 12 Yes, 10+ hours a week 275 2 15
Marital status* Self-rated health*
Married 5,570 50 21 Excellent 1,219 11 17
De-facto 1,470 13 10 Very good 3,931 35 19
Sep/Divorced 978 9 18 Good 4,058 37 16
Widowed 585 5 21 Fair 1,575 14 13 Single, never married
2,583 23 10
Poor 318 3 12
Highest education * Geographic location*
Bachelor + 2,449 22 23 Major city 6,865 61 15
Diploma 988 9 23 Inner Regional area 2,797 25 21
Certificate 2,376 21 15 Outer regional/ remote 1,525 14 22
Year 12 1,659 15 14
Year 11 or below 3,711 33 14 Life satisfaction*
Score <8 3,497 31 14
Employment status* Score 8+ 7,678 69 18
Employed F-T 4,935 44 13
Employed P-T 2,365 21 21 Importance of religion‡*
Unemployed 327 3 9 Of little/ no importance 6,195 56 14
Not in labour force 3,560 32 20 Medium importance 2,143 19 16
Important / very important 2,809 25 25
Total 11,187 17
1 Weighted by hhwtrps (responding person sample weight). *Difference in percentage volunteering is significant at p<=0.05, using chi-square test. ‡ Data on the importance of religion was from HILDA Wave 7 (2007) as the importance of religion was not asked in Wave 8 (2008).
20
Table 2. Logistic regression of volunteering, Wave 8
Total Males Females
Odds ratio Odds ratio Odds ratio
Sex
Male (ref) --
Female 1.10 Age group
15-24 0.73** 0.78 0.71* 25-34 0.83* 0.88 0.81 35-44 -- -- --
45-54 1.14 1.25 1.1 55-64 1.32*** 1.55*** 1.18 65+ 1.70*** 2.35*** 1.42**
Marital status
Married (ref) -- -- --
De-facto 0.49*** 0.45*** 0.53*** Separated/Divorced 0.84* 0.70** 0.97 Widowed 0.89 0.68 0.96 Single, never married 0.65*** 0.69** 0.64***
Country of birth
Australia (ref) -- -- --
Other English-speaking 0.93 0.88 0.97 Non English-speaking 0.63*** 0.47*** 0.75**
Child composition
No children at all -- -- --
No resident children aged under 15 1.04 1.21 0.93
Resident children aged only under 5 0.39*** 0.56** 0.27***
Resident children under 5 & also 5-14 1.21 1.34 1.04 Resident children aged 5-14 1.78*** 1.49*** 1.88***
Main carer of elderly/disabled person
No (ref) -- -- --
Yes, for <10 hours a week 2.26*** 2.89*** 1.96*** Yes, for 10 hours a week or more 0.80 0.59 0.91
Highest level of education
Bachelors+ 2.23*** 2.24*** 2.31*** Diploma/Adv. Diploma 1.69*** 1.60*** 1.85*** Certificate (ref) -- -- --
Year 12 1.22** 1.08 1.32**
Year 11 or below 0.86* 0.77** 0.93
Table 2 continues…
21
Table 2. Logistic regression of volunteering, Wave 8 (continued).
Total Males Females
Odds ratio Odds ratio Odds ratio
Employment status
Employed full-time (ref) -- -- --
Employed part-time 1.85*** 1.66*** 2.10*** Unemployed 1.28 1.21 1.39 Not in the labour force 1.77*** 1.27* 2.30***
Equivalised household income
Bottom third (ref) -- -- --
Middle 0.95 1.00 0.93 Top third 0.85** 0.85 0.86
Life satisfaction
Life satisfaction score <8 (ref) -- -- --
Life satisfaction score 8+ 1.04 0.99 1.08
Self-rated health
Excellent 1.34*** 1.33* 1.36** Very good 1.30*** 1.28** 1.32*** Good (ref) -- -- --
Fair 0.72*** 0.8 0.66*** Poor 0.61*** 0.84 0.47***
Geographic location
Major city (ref) -- -- --
Inner Regional Australia 1.27*** 1.37*** 1.22** Outer Regional /Remote Australia 1.66*** 1.71*** 1.61***
Number of observations 10,981 5,106 5,875
Prob>chi2 0.00 0.00 0.00 *** p<0.01, ** p<0.05, * p<0.1
22
Table 3. Event-history analysis of starting to volunteer, or breaking from volunteering
Volunteer start Volunteer break
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
Total Ages 15-34 Ages 35-54
Ages 55+
Total Ages 15-34 Ages 35-54 Ages
55+
Sex
Male (ref) -- -- -- -- -- -- -- -- Female 1.31*** 1.31*** 1.34*** 1.29** 0.81*** 1.03 0.68*** 0.83
Age group 15-24 0.63*** 1.14 1.46*** 1.20 25-34 0.65*** (ref) 1.35*** (ref) 35-44 (ref) 1.27** (ref) (ref) 0.76 (ref) 45-54 0.69*** 0.87* 0.93 0.89 55-64 0.72*** 0.98 (ref) 0.82** 0.62*** (ref) 65+ 0.46*** 0.73** 0.73*** 0.83
Marital status Married -continuing (ref) -- -- -- -- -- -- -- -- Cohabiting-continuing 0.63*** 0.65*** 0.72*** 0.85 1.08 1.03 1.40* 0.15** Start or end of marriage/cohabitation 0.96 0.92 0.69* 3.51*** 2.21*** 2.29*** 2.91*** 1.34 Single-continuing 0.79*** 0.92 0.76*** 0.88 1.05 1.09 1.12 0.93
Highest education level Bachelors+ 1.42*** 1.55*** 1.56*** 1.27 0.84** 0.93 0.83 0.78 Certficiate/Diploma (ref) -- -- -- -- -- -- -- -- Year 12 1.17** 1.20* 1.28** 1.01 1.02 0.95 1.26 0.86 Year 11 or below 0.71*** 0.79** 0.64*** 0.71*** 1.13 1.11 1.15 1.15
Employment transitions Continuously working (ref) -- -- -- -- -- -- -- -- Continuously not working 1.56*** 1.36*** 2.08*** 1.32** 0.89 0.78 0.88 0.91 Started or stopped working 1.70*** 1.48*** 2.13*** 2.15*** 1.74*** 1.91*** 2.15*** 1.01
Table 3 continues…
23
Table 3. Event-history analysis of starting to volunteer, or breaking from volunteering (continued).
Volunteer start Volunteer break
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
Total Ages 15-34 Ages 35-54
Ages 55+
Total Ages 15-34 Ages 35-54 Ages
55+
Equivalised income (thirds) Bottom third 1.03 1.11 0.94 0.96 1.01 0.92 0.84 1.39** Middle third (ref) -- -- -- -- -- -- -- -- Top third 0.86*** 0.96 0.89 1.08 1.01 0.87 0.99 1.34*
Health changes Improvement compared to one year ago 1.08 1.13 1.04 1.07 1.06 0.99 1.16 0.95 About the same as one year ago (ref) -- -- -- -- -- -- -- -- Worse compared to one year ago 0.74*** 0.82 0.82* 0.59*** 1.15* 0.96 1.20 1.23
Caring Not caring (ref) Caring <10 hours a week 1.94*** 3.99*** 1.62*** 1.36* 0.61*** 0.42*** 0.66*** 0.64* Caring 10+ hours a week 0.68** 2.10** 0.51** 0.51** 1.45** 0.34* 1.52 2.00**
Country of birth Australia (ref) -- -- -- -- -- -- -- -- Other English-speaking 0.91 0.72* 0.99 0.89 1.08 1.07 0.99 1.23 Other non-English speaking 0.68*** 0.88 0.59*** 0.53*** 1.26** 0.93 1.32* 1.47*
Moved in the last 12 months No (ref) -- -- -- -- -- -- -- -- Yes 0.84*** 0.79*** 0.90 1.27 1.58*** 1.20 1.56*** 3.05***
Geographic location Major city (ref) -- -- -- -- -- -- -- -- Inner Regional Australia 1.36*** 1.33*** 1.35*** 1.37*** 1.07 1.11 1.00 1.20 Outer Regional/Remote Australia 1.45*** 1.41*** 1.52*** 1.48*** 0.89 0.80 0.83 1.15
Table 3 continues…
24
Table 3. Event-history analysis of starting to volunteer, or breaking from volunteering (continued).
Volunteer start Volunteer break
Model 1a Model 2a Model 3a Model 4a Model 1b Model 2b Model 3b Model 4b
Total Ages 15-34 Ages 35-54
Ages 55+
Total Ages 15-34 Ages 35-54 Ages
55+
Time 0.95*** 1.01 0.94*** 0.89*** 0.94** 1.00 1.00 0.91
Birth in the last 12 months No (ref) -- -- -- -- Yes 0.52*** 0.51** 1.48 3.96***
Child composition
No children /no own resident children aged under 14 (ref) -- --
Own resident children only under 5 0.95 1.13 1.39 1.55 Own resident children under 5 & 5-14 2.74*** 2.08*** 1.03 0.65*** Resident children aged 5-14 but not <5 2.10*** 2.00*** 0.75 0.89 Number of respondents 11,836 5,434 3,900 2,527 2,900 735 1,330 832 Number of observations 47,081 21,217 16,263 10,532 8,988 1,811 4,309 2,862 Avg. observations per respondent 4 3.7 4.2 4.2 3.1 2.5 3.2 3.4
*** p<0.01, ** p<0.05, * p<0.1
